Skip to main content

TrainableClassifierCodeActivity Class

TrainableClassifierCodeActivity is an abstract class that implements the UiPath.DocumentProcessing.Contracts.Training.ITrainableClassifierActivity interface and the System.Activities.CodeActivity abstract class.

Definition​

  • Namespace: UiPath.DocumentProcessing.Contracts.Training
  • Assembly: UiPath.DocumentProcessing.Contracts

Description​

Abstract class that implements the UiPath.DocumentProcessing.Contracts.Training.ITrainableClassifierActivity interface and the System.Activities.CodeActivity abstract class.

This class should be implemented in order to construct a CodeActivity for classifiers training, compatible with the Train Classifiers Scope activity and the document processing infrastructure.

Training parameters in classifier activities​

All InArguments are auto-populated by the Train Classifiers Scope activity for all sub-activities. InArguments include the following information:

  • Information about the document that was classified, specifically original file path, text version, and document object model.
  • Details about each field that requires training, expected to be processed by the training activities. These types are defined by the Train Classifiers Scope activity in the Configure Training wizard.

This activity does not have any OutArguments.

Members​

Constructors​

TrainableClassifierCodeActivity() - The protected default constructor.

Methods​

  • GetCapabilities() - Gets, at design time, a list of document types the classifier is capable of classifying. This method will only be called at design time, and as such it should NOT use any InArgument(s).
  • ProvidesCapabilities() - Returns true if the classifier is able to provide a list of capabilities.

Properties​

  • DocumentPath InArgument<string> - The file path of the document.
  • DocumentText InArgument<string> - The text version of the document.
  • DocumentObjectModel InArgument<UiPath.DocumentProcessing.Contracts.Dom.Document>
    • The document object model of the document.
  • HumanValidatedData InArgument<UiPath.DocumentProcessing.Contracts.Results.ExtractionResult[]>
    • The extraction results for the processed document, that you want to use for training.
  • PersistenceId string - The unique, opaque ID that is used by the scope activities in order to correctly track and configure the training activity.
  • PredictionId <string>; - Identifies that the same document is processed in a workflow. It's composed of DocumentId and JobId.